Dr. Halima Bensmail: “The implementation of AI comes with great responsibility and ethical considerations.”

Engaging more women into the AI industry is very simple, but it requires the work and support of many. We need to include more women in conferences, workshops, and trainings, whether through part of panel discussions or in delivering presentations on recent work and research. We also need the parents of young girls to be […]

The Thrive Global Community welcomes voices from many spheres on our open platform. We publish pieces as written by outside contributors with a wide range of opinions, which don’t necessarily reflect our own. Community stories are not commissioned by our editorial team and must meet our guidelines prior to being published.

Engaging more women into the AI industry is very simple, but it requires the work and support of many. We need to include more women in conferences, workshops, and trainings, whether through part of panel discussions or in delivering presentations on recent work and research. We also need the parents of young girls to be more open and encouraging in diverse career paths, including engineering.

As part of my series about the women leading the Artificial Intelligence industry, I had the pleasure of interviewing Dr. Halima Bensmail.

Dr. Halima Bensmail is currently a principal scientist of Statistical Machine Learning and Computational Biology at Qatar Computing Research Institute (QCRI), an entity of Qatar Foundation, and holds a Joint Professor position at the College of Computer Science and Engineering at Hamad Bin Khalifa University (HBKU). She also serves as an advisory member of AIQT Foundation, an artificial intelligence and quantum technology center based in Davos, Switzerland.

Dr. Bensmail received her PhD from the University of Pierre & Marie Currie (Paris 6) in France, and was trained at the University of Washington in Seattle, the Data Theory Group at Leiden University (The Netherlands) and Fred Hutchinson Cancer Research Center in Seattle. She joined academia as an Assistant and Associate Professor of Statistical Machine Learning at the University of Tennessee and the University of Virginia.

Dr. Bensmail has published more than 90 peer reviewed papers in high impact journals, such as Nucleic Acids Research, Bioinformatics, Briefings in Bioinformatics, Journal of American Statistical Association and Pattern Recognitions to name a few. She teaches and advises several PhD students in the field of machine learning. She has also received several distinguished awards from The International Federation of Classification Society in Rome, Qatar Foundation Research Forum in Doha and The International Information Complexity and Statistical Modeling in High Dimensions in Turkey.

Thank you for joining us! Can you share with us the “backstory” of how you decided to pursue this career path?

My mother’s side was well-known for their affinity for mathematics. This definitely had a big impact on my skills, motivation, and over-the-top work ethic when it came to mathematics. Moreover, all of my teachers, both in primary and secondary schools, were always impressed by how well I did in math and consequently encouraged me to take this path despite my interest in biology.

At University, I was torn between pursuing a career in pure mathematics — as I was interested in teaching and research — and applied mathematics, an in-demand field at the time. Again, thanks to one of my professors, I was encouraged to pursue applied math since I had a real passion for changing the world through interdisciplinary technological innovations. That choice has allowed me to apply my mathematics skills to interdisciplinary research fields, starting from Management (University of Tennessee), Social and Behavioral Science (Leiden University), Public Health (University of Washington and University of Virginia) to Chemistry and Health Science (Fred Hutchinson Cancer Research Center).

What lessons can others learn from your story?

If you want to achieve a goal, always be persistent and keep trying no matter what. Always keep the right people around you because they’ll fuel your desire to accomplish that goal. My life is full of so many anecdotes that demonstrate the importance of these two things, which enabled me to keep pushing the limits.

One decision point particularly stands out to me. While a second year PhD student, I was working as a TA in a high school at Montgeron, which is a commune in the southeastern suburbs of Paris, France. After teaching there for two years as temporary teacher, I was offered a permanent position and a very attractive salary. While this security and stability was not without its perks, I thought if I accepted the offer, it would mean that I would become a high school teacher and quit my dream. I decided to decline the offer and go on with my third year of the PhD. That is when I met Professor Raftery, who invited me to spend few years in his lab, and this opened the door wide to the research world.

Can you tell our readers about the most interesting projects you are working on now?

I am working on some exciting projects in the healthcare field — one focused on accelerating the drug discovery process through structure-based drug design, the other on identifying effective methods for pre-diabetes screening that would help individuals reverse the development of pre-diabetes.

The design and optimization of a chemical structure with the goal of identifying a compound suitable for clinical testing — also called “structure-based drug design” — has emerged as a valuable pharmaceutical lead discovery tool for drug candidate. This field shows potential for accelerating the discovery process of a drug, while reducing developmental costs and boosting the potency of the drug that is ultimately selected.

In structure-based drug design, it is extremely valuable for us to have knowledge of the 3D shape of the drug, which helps guide the design of the best drug molecules. For a drug to bind strongly to its target (protein) and have a strong effect, it needs to be able to adopt the right shape to fit into the protein binding site. If a molecule has to dramatically change shape to bind to the protein, it is likely to bind poorly and therefore be unsuitable as a drug. In contrast, a molecule that already has the right binding is more likely to bind strongly to the target and be a good drug.

As can be imagined, this method heavily relies on the production of high-resolution, 3D structures of the drug target, obtained through a technology called “X-ray crystallographic analysis.” To overcome the expensive cost, high attrition rate and series of trial and error settings, many in-silico methods have been developed to predict the crystallization propensities of proteins. To help accelerate the process, I have developed with my team a machine learning algorithm (Insilco) based on a deep learning algorithm (a popular AI algorithm), which predicts the 3D structure of a protein and its crystallization. These two state of the art algorithms beat in performance most of the recently published algorithms in this area. These two softwares, named “DeepCrystal” and “Bcrystal,” are published at the Journal of Bioinformatics, and their webservers are publicly available online at!pages/home.

In eHealth, I’m working on developing a non-invasive, pre-diabetes screening score that can help the average person check their risks of developing pre-diabetes. The score can help individuals reverse the development of pre-diabetes before entering the later stages when it becomes too late. The standalone webserver to test pre-diabetes score is available at

None of us are able to achieve success without some help along the way. Is there a particular person who you are grateful towards who helped get you to where you are? Can you share a story about that?

I owe my gratitude to three special people for making me who I am today.

My mother, who encouraged me and pushed me to work hard in my path of research when I had a choice between going to France to continue my studies in Applied Mathematics or becoming a pilot — which would have been an entirely different course of life, though not without excitement!

My mathematics teacher, Mr. Addy, who encouraged me in middle school to become a mathematician. He was always telling us that there is only one way to get better at math –

practice. But, he knew that in order to practice, we must find math interesting enough. So we were given interesting puzzles, and he frequently used math to explain concepts of life to us. As a funny anecdote, he jokingly called me “the macaque of the atlas” to motivate me more and more.

And last but definitely not least, my professor, Dr. Adrian Raftery from the University of Washington. During a visit to University of Washington, he encouraged me to visit him and hosted me in his lab during and after obtaining my PhD.

What are the 5 things that most excite you about the AI industry? Why?

  1. The AI industry is a fascinating field because it is full of possible solutions with virtually limitless applications to solve problems. Even if many people are working in this space in top labs like Google Brain, DeepMind, SpaceX, or Tesla, anyone can enter the field quickly. If by chance one stumbles upon a novel and unique idea, they can be part of a solution to a problem that we encounter every day.
  2. Intelligence machines are influencing almost every aspect of our lives to help improve efficiencies and augment our human capabilities. AI is so intertwined with all that we do; to the point that we cannot imagine living life without it.
  3. The future of drug discovery by itself is interesting, but it’s also one of the most exciting applications of AI. Drug discovery traditionally comes with hefty costs and an extended discovery process. AI helps to solve these challenges by looking at millions of chemical compounds and available drugs to predict their ability to cure a wider array of diseases.
  4. Similarly, the AI industry is playing a major role in improving medical services and medical devices. When it comes to diagnostic medicine, wearable medical devices with AI provide a foundation for medical image analysis, gene sequencing, and biosensors for the rapid detection of bacterial infections in biological fluids. Biosensor is a culture-free diagnostic method utilizing nanotechnology, which needs to be implemented with machine learning and AI to be able to classify various bacteria. This technology is highly sensitive, fast, cheap, and can be implemented at bedside (hospitals) using portable (hand-held) Raman spectroscope. I am involved in one of these projects and my main contribution is to populate the little biosensors with a classification algorithm that can classify a set of spectra into one class of family of bacteria. Another concept that recently started attracting attention — especially for pharmaceutical companies — and which technology has finally caught up to is known as “generative model.” This is a type of product development in which algorithms generate multiple product designs (models) based on scientist-supplied parameters such as size and weight, distribution, raw materials used and the intended functionality of the product. From there, the algorithm provides suitable candidates for testing.
  5. AI has great potential to promote the ongoing development of medical diagnostic technology because of its excellent performance in the field of data processing beyond human experts.

What are the 5 things that concern you about the AI industry? Why?

  1. The implementation of AI comes with great responsibility and ethical considerations. We bear the burden of any sort of hardware malfunctions if AI is used incorrectly; and it’s important to highlight that AI algorithms make predictions based on the training they are given. For example, producing a human-like robot for which ethical data were not programmed, or what we call “robot’s morality” — engineers that design functionally moral robots confront many constraints. Any approach to building machines capable of making moral decisions in critical situations will have to be assessed in light of the feasibility of implementing the theory as a computer program.
  2. An erroneous algorithm, based on skewed or incomplete data, will always make incorrect and unfavorable predictions. Again, this goes back to the fact that AI algorithms make predictions based on the training they are given. Consider the use of AI in court to make predictive judgments. Now consider that AI is using data skewed heavily toward certain groups of people to make that judgment.
  3. Small and mid-sized organizations struggle a lot when it comes to adopting AI technologies, especially because AI adoption can be a costly affair.
  4. As the volume of data available for processing grows exponentially, the computation speed requirements will need to grow with it. It is imperative to develop next-generation computational infrastructure solutions to support the application of AI.
  5. There is a myth that AI can solve all kinds of problems, whereas AI is just a tool that can operate only with the indulgence of human brains. AI can definitely boost the outcome of something that will replace jobs typically held by humans, such as the automation of routine or common work.

As you know, there is an ongoing debate between prominent scientists, (personified as a debate between Elon Musk and Mark Zuckerberg,) about whether advanced AI has the future potential to pose a danger to humanity. What is your position about this?

AI is, and always will be, a part of our daily lives — it has made our lives a lot easier. We have witnessed the development of robotic surgery, programming nanodevices interacting with cells and tissues and carrying out very specific tasks, autonomous vehicles, flights, medical devices, facial and emotional recognition, just to name some examples.

But we need to think about the downside and risk now before it is too late. The implementation of AI application comes with great responsibility, and a malfunction or malicious programming can cost humanity so much. This sentiment has also been highlighted by influential figures such as Bill Gates, the late Stephen Hawking and author Nick Bostrom in his book, “Superintelligence: Paths, Dangers, Strategies.” They have expressed concern about creating machines that are more intelligent than we are. And it is a real possibility.

What can be done to prevent such concerns from materializing? And what can be done to assure the public that there is nothing to be concerned about?

One possibility is to create intelligence that does not harm humanity, which means we need to create super-intelligence with safeguards to ensure it will not get out of control.

Similar to our worry about the misuse of nuclear power, we need to encourage collaboration across countries and stakeholder groups around the opportunities and issues presented by AI. As humans, we also need to assist and interact with AI, so that when programmed and used and enforced to learn, it is using human prior information to guide it. This is what we call Bayesian and enforced learning in mathematics.

How have you used your success to bring goodness to the world? Can you share a story?

Aside from being a scientist, I am a teacher. I use class time with my graduate students to share my knowledge of AI technology and motivate them to look at the possibilities in the field.

I have also been a part of a book about women in AI technology. The book entitled “Women in Tech: Inspiration No Fairytales” will be shared with young female students to learn about my experience in a field that has been dominated by men. I hope the book encourages them to get involved in AI. The book is part of a project called #eSkills4Girls, which is a global initiative of the German Federal Ministry for Economic Cooperation (BMZ) and the German Agency for International Cooperation (GIZ) that has more than 30 flagship projects worldwide. It aims to improve the digital skills of girls/women in rural and peri-urban areas through two fields of action: (1) Developing innovative career guidance modules that encourage secondary school girls to choose ICT trainings (2) To improve the employability of young female students in vocational training centers who have participated in interdisciplinary ICT modules.

Additionally, I have used Omics big data and machine learning to propel the understanding of biological mechanisms and to discover novel diagnostics and therapeutics. I have used Qatar Biobank data to discover the biological biomarkers that can predict Type2 Diabetes, and have also produced a mobile app for identifying individuals with undiagnosed pre-diabetes and for promoting behavior change.

Last, but not least, I have helped — alongside my colleagues from QCRI and Columbia University -design a machine learning algorithm named Robust Gradient Boosting Machine (RGBM), whose goal is to identify the main regulators of the molecular subtypes of glioma cancer (a very dangerous type of cancer that occurs in the brain and nervous system). The use of this machine reveals the identity and biological activities of the master regulators driving the transformation of two subtypes of mutant glioma and two subtypes of wildtype glioma cancers. It is very useful in slowing down the mortality rate from this dangerous cancer.

As you know, there are not that many women in your industry. Can you share 3 things that you would advise to other women in the AI space to thrive?

  1. AI is a reflection of those who make the technology. As humans are naturally biased, so will be AI. I recently attended the “Grace Hopper Celebration For Women In Computing” and in one of the presentations, the presenter showed an example in which Google Translate corrected the sentences ‘he is a babysitter’ and ‘she is a doctor’ to ‘she is a babysitter’ and ‘he is a doctor.’ So, I think that women who are undecided about whether or not to enter the field should not be discouraged. We need more women in the field because that enables more diversity in the development of AI, leading to more diversity in products.
  2. I think that car GPS voices — Siri from Apple, Alexa from Amazon, and other smart home and responsive bots — have female voices due to the fact AI engineers in charge of these bots had some level of partiality when putting together the responses. In turn, you can also engineer and inspire the next generation of smart devices and AI applications that permeate our lives.
  3. When I was a junior scientist, I was always intimidated by the heavy male presence during a discussion or a conference, even though I had the right answers to questions in my head. As I matured and developed more confidence, I felt more comfortable speaking my mind and standing out with my opinions and perspectives — as every scientist should do. I advise other women to feel comfortable being one of the few women in the meeting, to feel empowered sharing their thoughts, and support other women in the space.
  4. We need to close the gender gap in all areas of technology, including AI. Only 13% of those working in machine learning are female, but all is not lost. By encouraging women and girls to enter the field, demonstrating the possibilities, mentoring and promoting them, I believe we can close the gender gap.

Can you advise what is needed to engage more women into the AI industry?

Engaging more women into the AI industry is very simple, but it requires the work and support of many. We need to include more women in conferences, workshops, and trainings, whether through part of panel discussions or in delivering presentations on recent work and research. We also need the parents of young girls to be more open and encouraging in diverse career paths, including engineering.

What is your favorite “Life Lesson Quote”? Can you share a story of how that had relevance to your own life?

I like Benjamin Franklin’s quote: “To succeed, jump as quickly at opportunities as you do at conclusions.” I believe that one should seize any opportunity that comes your way, and not shy away from the obstacles that you encounter.

How can our readers follow you on social media?

You can follow me on LinkedIn:

I would also encourage readers to follow QCRI and HBKU’s Twitter and Facebook accounts so they can stay up to date on our work, especially in AI technology.
Share your comments below. Please read our commenting guidelines before posting. If you have a concern about a comment, report it here.
We use cookies on our site to give you the best experience possible. By continuing to browse the site, you agree to this use. For more information on how we use cookies, see our Privacy Policy.